|Home | About | Journals | Submit | Contact Us | Français|
It is widely accepted that pregnancy constitutes a unique developmental event. Unprecedented intrauterine actions of angiogenesis, immunity, and neuroendocrine regulation are juxtaposed to mechanisms of senescence that enable fetal growth and protection. The suppressive and regulatory factors that facilitate healthy pregnancy are under investigation. In non-pregnant systems of infection and inflammation the cytokine interleukin-10 (IL-10) has been widely investigated due to its potential as a key immunosuppressant in response to a multitude of inflammatory events. In the context of pregnancy, IL-10 levels increase markedly in women during early pregnancy and remain elevated well into the third trimester immediately prior to onset of labor. The role of IL-10 during pregnancy as a suppressor of active maternal immunity to allow acceptance of the fetal allograft has been a point of study. Moreover, secretion of IL-10 by a diverse set of maternal and fetal cells has proven to aid in the orchestration of normal processes of pregnancy. Interestingly, some of the more profound findings regarding the actions of IL-10 during pregnancy have manifested from research that focuses on aberrant pregnancy outcomes as a result of inflammation, hormonal imbalances, or gene-environment interactions. This review focuses on the role of IL-10 as a facilitator of successful pregnancy both as an immune suppressive agent and mediator of cross-talk between the placenta and the decidua. Importantly, we utilize investigations into adverse pregnancy conditions to further elucidate the multifarious role of IL-10 at the maternal-fetal interface.
Interleukin-10 was first reported by Mosmann and colleagues under the name of cytokine synthesis inhibitory factor (CSIF) as a protein with the ability to inhibit the activity of inflammatory T helper 1 (Th1) type cells 1. Investigation led to the deduction that CSIF was the major factor that defined a difference between Th1 and Th2 type T-cells as CSIF skewed T-cell activity toward inhibition. Although originally defined as a product of Th2 cells, this cytokine has now been shown to be produced by a wide set of cell types, including both immune and non-immune cells 2. Reports also demonstrated that one mode of IL-10 regulation is through a feedback loop that curtails excessive inflammatory events. For example, when monocytes are activated with lipopolysaccharide (LPS) a dual cytokine response is induced where pro-inflammatory cytokine production is countered by production of IL-10 3.
IL-10 began to flood the literature as a prominent cytokine that works in an autocrine and paracrine manner in response to the inflammatory limb of the immune system to sequester over-activation of pro-inflammatory signals. The capacity of IL-10 as a suppressive agent was bolstered by evidence that Epstein Barr Virus (EBV) contained a genomic insert with homology to the human IL-10 gene. It is hypothesized that EBV acquired the hIL-10 gene through evolution as a means to increase anti-viral responses during host infection 4. Importantly, research also showed IL-10 could act as a growth factor for lymphoid and myeloid cells under certain conditions, indicating that IL-10 was not solely an immunosuppressant 5.
X-ray crystallography confirmed that IL-10 is an acid sensitive homodimeric protein. Genetic data demonstrates that IL-10 is encoded on chromosome 1 of both mouse and humans, and that mIL-10 and hIL-10 are fairly conserved in their amino acid sequences sharing ~73% homology. hIL-10 and mIL-10 span 4.7 kb and 5.1 kb chromosome regions, respectively, yet both active forms are encoded by a series of 5 exons 2. Recent reports provide evidence for genetically mediated regulation of IL-10 production. Although several polymorphic changes have been identified in the IL-10 gene promoter, three sites at the −1082 (G/A), −819 (C/T), and −592 (C/A) positions have been best characterized for their regulatory influence. Later in this review we report that multiple cohort studies show SNPs in the promoter region of the IL-10 gene may correlate with increased susceptibility to particular adverse conditions of pregnancy 6–10.
The IL-10 receptor is composed of two subunits IL-10R1 and IL-10R2, known members of the interferon receptor family (IFNR). Expression of IL-10R is reported on hemopoietic as well as nonhemopoietic cells 11. IL-10R1 is constitutively expressed on placental cytotrophoblasts 12. IL-10R1 is mainly necessary for the binding of the IL-10 protein while IL-10R2 is specific to initiate a signaling cascade. IL-10R2−/− mice behave like IL-10−/− mice indicating that the second subunit of the receptor is essential for IL-10 signaling. The most well described signaling pathway specific for IL-10 binding is that of the Jak/STAT pathway. Briefly, Tyk2 and Jak1 are recruited to the IL-10R1/2 complex. Upon Jak1/Tyk2-mediated phosphorylation of the cytoplasmic tail of the receptors, STAT3 is recruited and its phosphorylation allows homodimerization and entry into the nucleus for transcription of IL-10-regulated genes. Importantly, the STAT3 complex also induces transcription of the protein SOCS3 which triggers a negative feedback loop of IL-10 regulation by blocking subsequent phosphorylation of Jak1 11.
Several clinical observations regarding pregnancy implicate a role of an anti-inflammatory regulator such as IL-10 13. A significant number of women with rheumatoid arthritis (RA), an inflammation driven condition, consistently reported diminished symptoms during pregnancy. In contrast, women with systemic lupus erythematosus (SLE), an antibody driven autoimmune disease, presented with increased symptoms during pregnancy. Taken together, these reports supported the postulate that an anti-inflammatory milieu, perhaps dominated by IL-10, was amplified during pregnancy most likely as a mechanism of tolerance toward the fetal-allograft.
Initial studies of the role of IL-10 during pregnancy were done in mice. Murine decidual tissues harvested across the spectrum of gestation showed that IL-10 was produced in supernatants and peaked at gestational day (gd)12 14. Administration of recombinant IL-10 in abortion prone CBA/DBA mice significantly abrogated the incidence of spontaneous fetal loss 15. In placental tissue obtained from normal pregnant women, immunohistochemical analysis coupled with ELISA showed that IL-10 was produced in a gestational age dependent manner. Levels of IL-10 from first and second trimester placental tissues were significantly higher than levels found in third trimester tissues, suggesting that IL-10 is intrinsically down-regulated at term to prepare for the onset of labor programmed by production of an inflammatory milieu 16.
Further studies elucidated the crucial role of IL-10 at the maternal-fetal interface as placental and decidual tissue from first trimester missed abortions showed decreased IL-10 production when compared to control tissues obtained from first trimester elective terminations 17. Similarly, a comparison of placental tissue from elective caesarean (pre-labor) and placental tissue obtained post labor showed higher IL-10 production in pre-labor tissues. Importantly, high IL-10 production in pre-labor tissues correlated to low prostaglandin-2 (PGE-2) levels, whereas the opposite held true for post labor tissues 18. These data established IL-10 as a key contributor to the balance of pro versus anti inflammatory signals that orchestrate proper pregnancy outcomes. Figure 1 presents a contemporary view of temporal potential of IL-10 at different stages of pregnancy.
Ten years later the role of IL-10 in pregnancy as an immunosuppressive agent is solidified, and recent studies have focused on its mechanistic properties. Investigation of pregnant IL-10−/− mice, when housed in a pathogen free facility, showed that IL-10−/− litter size and term of gestation were similar to those observed in congenic wild type (WT) mice 19. Interestingly, IL-10−/− mice showed increased placental size and larger areas of maternal blood sinuses as compared to WT controls 20. These data elucidated a role for IL-10 as a mediator of placental growth and remodeling. It is noteworthy that extravillous trophoblasts from first trimester exhibit poor IL-10 production while expressing high levels of message for matrix metalloproteinase-9 (MMP9), implying that invading trophoblasts may temporally down-regulate IL-10 expression to maintain their invasive, not necessarily endovascular, potential 21.
The maternal-fetal interface is composed of trophoblast cells of fetal origin intermingled with specialized maternal lymphocytes, stromal cells, and endothelial cells that comprise the decidua. Here we highlight studies that have defined production of IL-10 by trophoblast cells and subsets of maternal uterine lymphocytes, and summarize recent literature that delves into the intricate network of cellular crosstalk that mediates this control.
A conundrum in the field of reproductive immunology is the presence of uterine natural killer cells (uNK) throughout the decidua. NK cells operate through the missing self hypothesis where lack of MHC antigen presentation on a target cell leads to activation of the NK cells and resultant cytotoxicity 22. In an organ that was once considered immune-privileged, it is difficult to rationalize the presence of NK cells. However, it has been postulated that the expression of non-classical MHC type I molecule HLA-G on trophoblasts, particularly those with extravillous differentiation, plays a regulatory role in controlling NK cell cytotoxic activation 23. Interestingly, IL-10 has been shown to induce HLA-G on trophoblasts 24. HLA-G is present in different isoforms and has become a focus of an intense debate for the exact role that it plays.
While the mechanisms of HLA-G based antigen presentation remain to be fully elucidated, the role of IL-10 as both a paracrine and autocrine regulator of trophoblast activity is apparent. Although villous cytotrophoblasts produce IL-10, it is not clear how trophoblast differentiation and invasion are controlled at this level. IL-10 decreases MMP9 transcription in villous cytotrophoblasts 21, 25. This could be one mechanism by which cytotrophoblasts are selected for further differentiation and invasion. Compared to villous cytotrophoblasts, extravillous trophoblasts are intrinsically poor in IL-10 production, thus allowing MMP expression and invasion competency. This may require paracrine activity within the placental meshwork 12, 26–28. It is noteworthy that simultaneous activity of progesterone and IL-10 on trophoblasts may work to sequester pro-inflammatory responses and to enhance regulated cross-talk between the placenta and the decidua 29.
Uterine NK cells and monocytes are two major immune populations present in the decidua during pregnancy. These cells have the ability to produce, and to be regulated by IL-10 30–31. Importantly, the characteristic immune functions of these cells, normally identified as cytotoxic killers, are altered in the context of pregnancy where their ability to aid in angiogenesis and placental regulation is paramount. The role of uNK cells in placental growth is discussed later in this review in the context of functional studies undertaken in our lab.
Several human studies lend evidence to the regulation of uNK cells or monocytes by IL-10. First trimester tissue from human surgical abortions was obtained and lymphocytes were isolated. IL-10 production was assessed in comparison to peripheral blood mononuclear cells (PBMCs). Baseline IL-10 production from uterine monocytes and uNK cells was significantly elevated above PBMC production. Furthermore, stimulation with LPS of these cells enhanced production of IL-10, indicating that a pro-inflammatory stimulus can elicit a suppressive cytokine response in the context of the uterine milieu 18, 32. Finally, primary human uterine monocytes were isolated from decidual tissues obtained post labor, and pre-labor (caesarean) and production of IL-10 was measured via ELISPOT. IL-10 from pre-labor tissue was markedly increased above post labor levels, and this correlated to an increase in COX-2 mRNA signal in post, but not pre, labor tissues 33. These findings highlight the necessity of inflammatory signals to induce labor that couple to mechanisms aimed at silencing the action of IL-10.
Important insights into the immunological capabilities of uNK cells and decidual monocytes at the maternal-fetal interface have come from a mouse models of pregnancy established in our lab and others. We have studied IL-10−/− mice and their WT counterparts for pregnancy outcomes in response to exposure to inflammatory agents on gd6 or gd14, to mimic early pregnancy loss or preterm birth, respectively. Briefly, toll-like receptors (TLRs) are a group of innate immune receptors that recognize different pathogenic motifs. Injection of various TLR agonists at different gestational ages mimics maternal infection and allows for assessment of adverse pregnancy outcomes due to dysregulation of decidual immunity in the presence or absence of IL-10.
Studies with LPS, a TLR4 agonist, in IL-10−/− and WT mice induced fetal resorption (FR) or preterm birth on gd12 or gd17, respectively. Importantly, we found that IL-10−/− mice were highly susceptible to low doses of LPS, but WT mice required at least a 50 fold higher dose to induce adverse pregnancy outcomes. Dysregulation of innate immunity was similar in IL-10−/− and WT mice in that uNK cells became cytotoxic, produced TNF-α, and infiltrated the placental zone 19, 34. Similar results were observed in response to TLR9 agonist CpG. Negative pregnancy outcomes were induced in IL-10−/− mice at very low doses of CpG whereas WT mice responded negatively only when doses were increased 20 fold. Importantly, investigation of the cellular immune dysregulation showed that macrophages, not uNK cells, were activated to produce TNF-α and infiltrate the placental zone 35. Taken together, these results demonstrate that in response to certain pathogens, IL-10 is a protective agent. Furthermore, the absence of IL-10 allows investigation of the pathogenesis of bacterial and viral motifs at sub-clinical levels. On the other hand, as a simple rule of nature, IL-10 can not be presented as a global suppressive agent against all infectious agents. Our recent results are intriguing in that IL-10 does not protect pregnancy against mimics which represent double strand RNA viruses (unpublished observations).
T regulatory (Tregs) cells in the decidua have recently come under the microscope of pregnancy research. Their characteristic ability to produce suppressive cytokines in response to foreign antigen makes Tregs promising therapeutic targets for intervention toward adverse pregnancy outcomes. Tregs are characterized as CD4+/CD25+/Foxp3+ and their ability to produce IL-10 is well documented 36. The presence of Tregs was assessed in the murine decidua. Unpublished data from our lab and others show that murine Tregs appear in the estrous cycle and increase early in pregnancy, peaking on gd10-12 and declining thereafter 37, 38. Spontaneous fetal resorption in abortion prone CBA/DBA mice can be abrogated by adoptive transfer of Tregs harvested from same gestational age WT mice. Importantly, neutralization of IL-10 in the aforementioned experimental setting abolishes the ability of WT Tregs to rescue CBA/DBA fetal resorption 39. Finally, recent observations in humans have shown that decidual Tregs can inhibit immune stimulation of conventional T cells through cell-cell contact or IL-10 production 40.
Recent findings suggest that uterine Tregs may be of peripheral blood origin and their development toward the uterine phenotype may be under hormonal control 41. Migration studies with human decidual Tregs show that Tregs migrate to areas of hCG production. Women with ectopic pregnancies or spontaneous abortion show decreased IL-10 production coupled to low levels of Treg migration to trophoblast/hCG+ dense regions 42. Interestingly, murine CD4+/CD25− cells treated with E2 were converted to Foxp3+ T cells which produced IL-10 lending further evidence that Tregs may be under hormonal control 43. However, one study posits that decidual Treg development may be driven in part by the presence of paternal antigen as pseudopregnant females (mated with vasectomized males) showed increased levels of decidual Tregs 44. Unpublished data from our lab shows that Treg numbers do not differ between WT and IL-10 null pregnancies over the spectrum of gestation. However, we have begun to address differences in functionality of Tregs from IL-10−/− versus WT mice. Since Tregs from IL-10−/− mice are deficient in this cytokine, it is tempting to speculate that these Tregs may elicit a distinct regulatory role, particularly in response to inflammatory triggers. These studies may lend promising insights to Tregs as therapeutic targets due to their ability to influence pregnancy outcome through IL-10 dependent or independent mechanisms.
While specific decidual cell subsets still remain to be characterized, the role of IL-10 is manifesting from breakthrough work regarding crosstalk between different decidual immune cells. Recent research shows that gd12 murine trophoblasts co-cultured with dendritic cells (DCs) induced uNK cells to expand and produce IL-10, demonstrating that uNK cells are a rich source of IL-10 which could be required for maintaining their non-cytotoxic phenotype 45, 46. These data reveal that production of IL-10, and other pregnancy based cytokines, is context dependent and regulated by an intricate network of cellular cross-talk based on the decidual milieu. This assertion is further supported by a recent report that explored the role of Galectin-1, an immunoregulatory glycan binding protein, in the context of pregnancy. Gal1−/− mice displayed increased rates of fetal loss as compared to WT counterparts. Injection of recombinant Gal-1 into Gal-1−/− mice rescued pregnancy. This was directly associated with an increased number of decidual tolerogenic DCs which in turn induced expansion of IL-10 producing Tregs. Importantly, IL-10 neutralization or Treg depletion upon Gal-1 reconstitution abrogated the rescue of pregnancy 47. Such a scenario could also be envisioned for human pregnancy (Figure 2). These data show the existence of an intricate network of trophoblast-DC-IL-10-Treg based fetal-tolerance which remains to be further elucidated.
Successful pregnancy outcome is associated with immune tolerance and de novo angiogenesis at the maternal-fetal interface. Is there a link between these two events and does IL-10 contribute to angiogenesis? Our recent work provides evidence for both these processes. We have demonstrated that the non-cytotoxic phenotype of human uNK cells is maintained through production of VEGF C by these cells and VEGF C-mediated MHC class I expression on endothelial cells and trophoblasts 48, 49. Interestingly, IL-10 was found to induce VEGF C production by first trimester trophoblast cells under certain conditions (unpublished observations). Along similar lines, our recent results invoke the role of the water channels aquaporins (AQPs), particularly, at the maternal-fetal interface. AQP1 is a potent effector of fluid volume regulation and is expressed in both human and mouse placenta. AQP1 plays an important role in angiogenesis and our recent work demonstrates that expression of the AQP1 channel may be directly controlled by the presence of IL-10. We show that IL-10 induces expression of aquaporin 1 (AQP1) in human trophoblasts as well as in murine placental tissues. Treatment with polychlorinated biphenyls (PCBs), ubiquitous environmental toxicants, abrogated this effect in IL-10−/− mice and induced preterm birth in response to the loss of AQP1. However, administration of recombinant IL-10 to PCB-treated IL-10 null mice restored expression of AQP1 and led to term pregnancy 50. Our unpublished data also suggests that IL-10 down-regulates the Notch-dll4 axis, a nemesis of angiogenesis. As a corollary to our results, tumor cells are known to produce IL-10. It is tempting to speculate that IL-10 production by tumor cells programs their escape from immune surveillance and promotes angiogenesis 51. Taken together, these observations warrant a thorough analysis of IL-10 and aquaporins as angiogenic factors at the maternal-fetal interface.
There have been several studies that couple IL-10 deficiency to adverse pregnancy outcomes such as recurrent spontaneous abortion (RSA), preterm birth, and preeclampsia. The mechanisms that may lead to poor IL-10 production at the maternal-fetal interface are not well understood. However, polymorphisms in the IL-10 gene promoter have been associated with dysregulated IL-10 production and several diseases. Recent studies have identified five single nucleotide polymorphisms (SNPs) at −3575, −2849, −1082, −819, and −592 positions in the human IL-10 gene promoter52–55. Similarly, the molecular effects of these SNPs in the IL-10 gene promoter remain to be elucidated in the context of pregnancy complications. Below, we provide a discussion on the association of IL-10 dysregulation and adverse pregnancy outcomes.
Preeclampsia occurs in 5–10% of pregnancies worldwide and is a systemic disorder resulting from poor placentation. Although the pathogenesis of preeclampsia remains poorly understood, defective trophoblast invasion and spiral artery remodeling are thought to induce placental ischemia/hypoxia which eventually results in production of inflammatory molecules 56. Systemic presence of inflammatory molecules or dysregulation of essential proteins may then cause the maternal syndrome diagnosed by elevated blood pressure, proteinurea, kidney pathology, and edema 57. Does reduced production of IL-10 contribute to poor placentation and induction of inflammatory molecules? Curiously, evaluation of placental tissue and serum samples from preeclamptic women has suggested reduced IL-10 production 58, 59. Serum samples from preeclamptic women disrupt endovascular interactions between trophoblasts and endothelial cells and lead to the full spectrum of preeclampsia-like features in IL-10−/− mice compared to WT mice (our unpublished observations). In this regard, research that links low levels of IL-10 coupled to decreased numbers of Tregs with this elusive disease of pregnancy may shed light on its causative agents 60.
Based on our recent results, we surmise that IL-10 reconstitution prevents onset of preeclampsia-associated features in both in vivo and in vitro models of preeclampsia. There may also be a genetic link to preeclampsia and analysis of SNPs in the IL-10 gene promoter is likely to provide insights into the nature of this disease. Interestingly, one genotype, −2849AA, is thought to be associated with a three-fold reduced risk toward acquisition of preeclampsia 61
Recurrent spontaneous abortion has been linked to an increase in CD56+ cells as well as an increase in TNF-α 62,63. However, the balance of this inflammatory cytokine may be skewed due to a lack of IL-10 production. PBMCs from women with RSA show increased cytotoxicity due to high levels of TNF-α, but levels of IL-10 production are significantly lower than control PBMCs 64, 65. Similarly, PBMCs from women with RSA show lower production of IL-10 upon stimulation with trophoblastic antigen as compared to normal pregnancy controls 66. We have previously demonstrated that decidual and placental tissue from spontaneous abortions showed reduced presence of IL-10 with no effect on IFN-γ compared to tissue from elective terminations 17. Thus, poor IL-10 production coupled with increased production of inflammatory molecules may be a trigger for early pregnancy loss or preterm birth. Furthermore, placental explants obtained from women undergoing preterm labor showed poor IL-10 production coupled to elevated prostaglandin release as compared to normal pregnancy control samples 67.
Based on these observations, we established mouse models for fetal resorption and preterm birth using IL-10−/− mice. As was aforementioned, our data are significant in that low doses of inflammatory triggers cause fetal loss or preterm birth depending on the gestational age dependent exposure to the trigger 19, 34, 35. These pregnancy complications are strongly linked with immune programming in the form of cytotoxic activation of uterine NK cells, macrophages, or T cells and TNF-α production depending on the nature of the inflammatory trigger. These results provide impetus for further investigation into the nature of infection/inflammation and the ensuing immune responses in both mouse models and humans.
It is well accepted now that IL-10 influences immune responses in a variety of ways. In the context of pregnancy, we propose that IL-10 exerts profound effects on linking immunity, angiogenesis, and maintenance of expression of molecules regulating fluid volume across the placenta. Our work in IL-10 null mice for the first time provides important clues to the pathogenesis of fetal loss, preterm birth, and preeclampsia. These observations have given rise to the hope that IL-10-based therapy may some day become a reality for enigmatic pregnancy maladies.
We would like to thank Tania Nevers for insightful critique and reading of the manuscript. This work was supported in part by grants from NIH and NIEHS, P20RR018728 and Superfund Basic Research Program Award (P42ES013660). This work was also supported in part by the Rhode Island Research Alliance Collaborative Research Award 2009-28.